Objective: The metabolic syndrome (Mets) is a clustering of cardiovascular disease risk factors: visceral obesity, hyperglycemia, hypertension and dyslipidemia. However, the criteria and definitions of Mets vary among countries, and medical evaluation is not unified as yet. The aim of the present study was to develop a Mets simulator as a tool for health policies, calculating prevalence corresponding to several diagnostic criteria based on checkup results for a group.
Methods: 1) The data for construction of the simulator: To investigate the characteristic of each diagnosis criterion component (abdominal waist circumference, BMI, triglycerides, HDL cholesterol, blood pressure, fasting blood sugar, and HbAlc) actual checkup data were applied to examine for a suitable normal transformation method. Next, variance-covariance matrices were obtained using transformed variables. 2) Construction of the simulator: multivariate normal random values were generated using e mean values input by the user and the variance-covariance matrices calculated from the above process. Next, Mets was diagnosed according to the five diagnostic criteria (NCEP-ATPIII, AHA/NHLBI, IDF, Japan, and Ministry of Health, Labor and Welfare) to obtain the prevalence for each criterion. By repeating the process, average and variance values for prevalence were calculated for each.
Results: The Mets simulator was constructed using checkup data in Tokyo and performance assessment and sensitivity analyses were accomplished. Next, two examples referred from the studies in Japan and America, both including missing diagnostic criterion components, were examined using the simulator. Mean and variance values of the prevalence, and missing values, were estimated using the simulator. Change in prevalence with change of values for diagnosis criterion components was calculated. The result clarified a part of the features of the diagnostic criteria such as similarities in those applied in Japan and IDF.
Conclusion: A simulator corresponding to the five diagnostic criteria of the Mets could be developed by generating multivariate random values, allowing assessment of Mets prevalence. Using the simulator, changes in prevalence according to changes in values of components of the diagnostic criteria were examined. Th simulator may bring useful information for health policies to prevent lifestyle related diseases.